DMS Statistics and Data Science Seminar

Speaker: Luca Insolia, Sant'Anna School of Advanced Studies

Title: Parasitic Mites, Pesticides and Extreme Weather Linked to Honey Bee Loss: A Study Across the United States Through Multiple Open Data Sources

Abstract: Honey bee (Apis mellifera) loss is widespread across the globe and has tremendous economic and biological impacts. To date, various local level studies attempted to explain this loss and bring forth proposals to invert the trend. We contribute to this line of research through a large-scale study over the Contiguous United States (CONUS), where we aggregate national level open-source databases that comprise information on different potential risk factors (e.g., weather, pesticides, parasites, mites, etc.) across different areas and periods. Specifically, we analyze quarterly data which cover most of the CONUS territory for the years 2015-2019 and we combine publicly accessible data sources on honey bee status and stressors, weather data, and land use. The different spatio-temporal resolution of these data is addressed through an effective up-scaling approach that generates additional statistical features by taking into account the distributional characteristics and interactions of the original variables – which significantly improves modeling performance. Treating this expanded set with state-of-the art feature selection methods we find strong evidence confirming, nation-wide, the role of Varroa destructor and pesticides in aggravating honey bee loss that was already observed in local studies. In addition we find that, even when controlling for other factors, extreme temperature and precipitation events are significantly linked to honey bee loss at the CONUS level.